When I started editing podcasts, it was normal for everyone to be in the same room. One of my clients traveled up and down the country, interviewing people using a couple of lapel mics plugged into a handheld recorder.
Of course, many shows in America were still recorded over the internet, but it was much harder back then. You had to ask each person to record their side of the call, sync up the tracks, and handle something called drift.
Kids today don’t know how good they’ve got it. Anyway, enough of the “old man yells at cloud” shtick. The point is, it’s become much easier to edit remote conversations, but some of the education has lagged behind.
This has led to many podcasts with sloppy editing that can be easily avoided. We’re talking about stuff that’s mostly not difficult to do and will transform your podcast from two people having a nice conversation to something that gets to the heart of what your listeners care about—and saves them time.
The two-part edit
I recommend editing in two passes: first, a technical edit, then a content edit. I use Descript, and when I’m editing conversations, I do the technical edit in the sequence view, where I can see the waveforms for each person speaking. I do the content edit in the script view, where I can see all the words.
If you’re editing a one-person show or you have everyone’s voices on one track, some of this stuff you just won’t be able to do. That’s why it’s always important to get everyone’s voices as separate audio files.
Okay, so what exactly should we remove, and in what order?
Crosstalk
This is when two people are talking at the same time, usually caused by lag. It happens all the time, but the listener doesn’t need to hear it.
Either one person will dominate and keep talking – in which case we can remove the other person’s speaking – or both will jam themselves in the doorway at the same time. We don’t need to hear the back-and-forth of “No, after you” – we can just remove that awkwardness and make it sound more fluid.
Active listening noises
When I get a new episode, I go through the sequence editor in Descript and remove anything where the person isn’t speaking. If it sounds like an active listening noise, I remove it. If it’s a laugh, I’ll leave it in. What I don’t need are noises like “Hmm”, “Oh yeah”, “That’s interesting”, “Oh”, “All right”, “Cool”.
When you can see the other person’s face, these noises make sense, and they’re natural. When I’m recording my own shows with a guest, I use big facial expressions to show I’m engaged, rather than making those noises. It just saves me the editing job. But if that’s how you communicate that you’re listening, carry on doing it.
From a listener’s perspective, though, we don’t need it because it pulls focus. It’s slightly distracting from what the other person is saying. So it’s fairly trivial to go into the sequence editor and remove those.
Conversational admin
This is stuff you’ll find at the beginning and end of the conversation. I’m sure you already remove these bits, but as a quick reminder, we don’t need “Can we hear you? Can you hear me? Is everything okay?” We don’t need too much of the goodbyes at the end of the call either.
You’ll probably have an outro at the end of your episode anyway. All you really need is a simple “Thank you so much for being on the show”, and the guest saying “It’s been a pleasure”.
What you probably carry on doing – because you’re human – is continuing with “That was great. Really great recording. Thank you. I really appreciated it”, and so on. All those things we do as human beings don’t need to be in the final cut. Yes, a podcast is a conversation, but not a real one.
Periods of silence caused by Internet lag
There’s a difference between pausing for thought or emphasis, and the lag after someone’s asked a question. If it’s longer than a second, it’s too long, unless you’re really giving someone time to think about the question.
Or think about it this way. If the silence would feel natural for two people in the same room, then leave it in. Otherwise, rip it out!
There are options within Descript that can help you do this in a couple of clicks, or you can easily do it when you’re going through your content edit.
If someone’s telling a particularly impactful or emotional story, you want to keep those emotional beats as they’re part of the story. You might occasionally want to keep a pause if someone’s going to give a particularly difficult answer to a question.
If they’re sharing something revelatory, that pause has value – it builds suspense. But if someone’s just stroking their beard, having a little think, we don’t need to hear that.
Small talk
Similar to conversational admin, small talk is usually more about the weather or providing a space to land. You want to make your guest feel settled, and to settle yourself. We just don’t need to hear them as listeners. So get rid of all that small talk. Be brutal.
People don’t listen to podcasts because they are genuinely candid conversations. That’s one of the things we tell ourselves. What they are are heightened versions of conversations. They’re the conversations without the bits we wish we didn’t have to endure as human beings sometimes.
I know sometimes small talk is nice, but it’s just getting us to the meat of a conversation. You need that ramp-up. But as listeners, we’re not actually there in the pub or the coffee shop or the corridor of a conference center with you. We don’t need to fill up in that sense. We just want to hear what you’ve got to say, and we want to spend that time productively. So a great way to respect people’s time is to eliminate that kind of small talk.
False starts
Often, people will start a sentence and then have a quick change of heart, think about it, and start again. You’ve got an opportunity there as you’re listening through to rip them out. Get rid of them, and it makes everyone sound smarter.
This is part of what I call IQ editing. It’s part of that process of making everyone sound their wittiest and sharpest, and it doesn’t take much effort. The listener won’t notice, which means you’ve saved them time. Compound that over a few episodes, and you could save them hours (really, it adds up).
Follow-up questions
When we’re asking questions we’re not so confident in, we often end up with something like this: “So, what’s your favourite colour? I mean, when we talk about colour, there are all sorts of different options. So I was just wondering, what’s… um, yeah. What’s your, what’s your… Yeah, what’s your favourite colour, I guess? Yeah.”
It feels really strange to read that as a piece of text, but it’s more common than you might think to find it in interviews.
If you want to clarify the question on the call, by all means, do that. But we don’t need to hear it because we can make the guests sound a little telepathic and a little smarter if they’ve somehow anticipated the question without you having to ask it.
It’s another one of these subtle little things we can do. The listener thinks, “Wow, that was an insightful answer to a simple question”. What they don’t know is that you asked a fairly comprehensive question which allowed the guest to expound.
But because they haven’t heard that, they’ve just heard a simple question, they think, “Wow, the guest gave a really comprehensive, expansive answer to that simple question. Aren’t they great?”
Umms
Remove umms as much as you can without interfering with the flow of the dialogue. Sometimes words flow into each other or the pause has meaning as I’ve already mentioned. In that case, leave them in.
But some people really do seem to say “um” every couple of words. Maybe they’re nervous, maybe they’re just not that fluid a speaker.
If you can spare the time and if it feels natural — if it doesn’t disrupt the flow of how they’re actually speaking — then if you’ve got the time to get rid of umms, do it.
But don’t obsess over it too much. And don’t rely on Descript’s automatically filler word removal tool unless you’re going to do a second-by-second listen, because it’ll mess up 50% of them.
Repetition
Often you might find a guest summing up their point at the end of a statement. This isn’t a hard-and-fast rule but I tend to remove it, as it’s just repetition that isn’t all that helpful. Often it’s just a way for the guest to signal that they’re reaching the end of their point, or a way for them to collect and clarify their thoughts.
Boring bits
If you’re ahead in your publishing schedule, a great thing to do is edit your episode and then give it a few days to rest. Pop it in your phone, and have a listen to it as you would any other podcast, for example while you’re doing the dishes.
Try not to be super attentive, but notice when you drift off. Notice when you’re thinking about something else.
It’s a little bit like mindfulness: knowing when your mind has drifted, and then gently guiding it back on track. In your case, it’s about noting that you’ve drifted, and asking what you can do to pull the listener’s attention back. That could be a little bit of music – a snatch of your theme tune – to help create a clean break between segments and give us a quick nudge if our attention is wandering.
Off-topic stuff
You might find stuff in your episode that just isn’t entirely relevant to the topic. Maybe it’s really interesting, maybe it feels like background colour, but a lot of the time you want to aim for shorter and more pacy. Again, this is about respecting the listener’s time, and making good use of it.
What if you’re on a time crunch?
If your time’s limited, but you still want to put out an episode that will keep listeners coming back, here are the three things I’d recommend you remove:
- Crosstalk and active listening noises
- Gaps due to internet lag
- Conversational admin and small talk
Add your response